A Fully Automated Two-Stage Segmentation Approach for Late Gadolinium-Enhanced Cardiac Magnetic Resonance Images in Personalized Cardiac Modeling
Computing in cardiology(2023)
摘要
Accurate automatic segmentation of LGE-CMR images is vital for personalized cardiac modeling. We developed a two-stage method: a DL-based solution for left ventricular (LV) segmentation and an MGMM solution for infarct tissue (IT) in LV, enabling fully automated cardiac segmentation. Ventricular models were constructed for three patients using segmented LGE-CMR images, and programmed electrical stimulation induced VT. Our method achieved an 81.21 DS for LV and an 82.9 DS for IT. Simulation results for these patients matched manual methods, indicating the efficiency and reliability of our two-stage approach for personalized cardiac modeling.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要